From: Jonathan Hanson <[EMAIL PROTECTED]>
Subject: Re: gEDA: Symbol inheritance test
Date: Sun, 11 Dec 2005 21:41:19 -0600
Message-ID: <[EMAIL PROTECTED]>

Jonathan,

> I don't know that we have to build a complete 'heavy' library at first.  
> As long as a strong tradition of submiting custom symbols when you make 
> them for your projects exists, the heavy library will grow on its own.  

Yes, but the point that Rick and I have been trying to convey over the years
is that the architecture of the library needs to be such that it can handle and
be manageable when it gets big. In addition to that there is several steps in
the actual workflow which comes with large designs (which not necessarilly
means large library) that might require a little different attitude.

> I personally had only planned to do a heavy set for basic passive 
> devices (capacitors and resistors) and then a few odds and ends like a 
> handful of transistor models and some diodes.  That covers most of what 
> I'll be using it for, and when I need some IC stuff later, I'll just 
> create the models as I go.

For a handfull of symbols (as compared to a huge library) you can do it either
way, it is when you grow out of that and need more, and more, and more... that
you are in trouble and start thinking various forms of reuse, parametrisation
etc. etc.

To comment further on John's comment, some of the proposals have been trying
to find methods by which a large amount of re-use can be utilized such that
more variants can be handled with less effort and less space. It comes at a
cost of added complexity. The benefit is however that a larger library can
quite easilly be created in terms of supported types.

Anyway, the resistance have been great for such thinking and I think it shows
where most of the guys doing the work (do keep up the good work!) is standing
on the issue.

Cheers,
Magnus

>  - Jonathan Hanson
> 
> Richard G. Munden wrote:
> 
> > The problem is solvable but it requires a change to the library 
> > architecture - a subject that has been discussed several times in the 
> > past.
> >
> > Rick Munden
> >
> >
> > John Doty wrote:
> >
> >>> I think the symbol inheritance should work and could be a great
> >>> addition, and the heavy libraries built could be configured in the
> >>> search path library, so both light and heavy symbols guys will be
> >>> happy.
> >>
> >>
> >>
> >> I don't think it's possible to satisfy the heavy symbols advocates. It
> >> would require many millions of symbols to cover every component on the
> >> market. As far as I'm concerned, the only sensible thing is to build 
> >> up a
> >> library of custom symbols representing the parts stock you are going to
> >> use for the design.
> >>
> >> I could wish that symbol directories specified in ./gafrc show up before
> >> the standard libraries in the menu.
> >>
> >> John Doty          "You can't confuse me, that's my job."
> >> MIT-related mail:                       [EMAIL PROTECTED]
> >> Other mail:                             [EMAIL PROTECTED]
> >>
> >
> >
> 

Reply via email to